Observation, control and maintenance of physical condition of sows in acceptable levels is critical to maintain animal welfare and production standards. Early recognition of animals that present atypical physical condition is important to prevent production losses. Currently, classification of body condition is done by subjective methods, thus is dependent on the opinion of the manager, which can generate differences between ratings. As alternatives to these subjective methods of classification, various methods have been proposed to obtain a more objective measure. Knauer & Baitinger (2015) developed a calliper that quantifies the angularity from the spinous process to the transverse process of a sow's back and concluded that this instrument can be used as a tool to standardise this classification. Another way to standardise this measurement would be to automate the process by analysing images generated by depth cameras. The present work aimed to obtain sow's body condition score (BCS) using a commercially available depth camera. This was done by correlating the scores obtained with a BCS calliper (scores ranging from 1 to 29) with the sow's body widths acquired from depth images. A multiple linear regression was performed with an R2 of 0.61, a standard error of 1.36, and average absolute error of 8.01% (1.05 units). These errors may be associated with poor repeatability (human error) and/or the measurement calliper. It is considered that there is a need for a reliable gold standard and depth images analysis could be a possible option.